Project Ender 3 V2
An Electrical, Software and Hardware Overhaul
Created By
Ervince Allen
Completed: June 2025
Version: 1.0
Email: contact@ervince-engineering.com
i Table of Contents
Table of Contents
Portfolio Summary ............................................................................................................. 1
Core Skill Highlight ......................................................................................................... 1
1. Introduction ................................................................................................................... 2
2. Print Head & Cooling System ........................................................................................... 3
Overview ........................................................................................................................ 4
Ducting .......................................................................................................................... 4
Overview .................................................................................................................... 4
Pre-CFD ..................................................................................................................... 5
Post-CFD .................................................................................................................... 5
Part Cooling Fan Efficiency .......................................................................................... 5
ADXL345 Mounting ......................................................................................................... 6
Overview .................................................................................................................... 6
More Information ........................................................................................................ 6
Larger Fan Mount ............................................................................................................ 7
Overview .................................................................................................................... 7
Reflections ..................................................................................................................... 7
Starting from Scratch................................................................................................... 7
Future Improvements .................................................................................................. 8
3. Dual Z-Axis Synchronisation ............................................................................................ 8
Overview ........................................................................................................................ 8
Solution ......................................................................................................................... 9
Reflections ..................................................................................................................... 9
Tensioner Issue ........................................................................................................... 9
Future Improvements .................................................................................................. 9
4. Filament Drying System ................................................................................................ 10
Overview ...................................................................................................................... 10
Initial Solution .............................................................................................................. 10
Solution ....................................................................................................................... 10
Reflections ................................................................................................................... 10
Thermal Performance ................................................................................................ 10
ii Table of Contents
Ventilation Limitations ............................................................................................... 10
Temperature Stratification ......................................................................................... 10
Current Material Limitations ...................................................................................... 11
Future Improvements ................................................................................................ 11
5. Host & MCU ................................................................................................................. 11
Overview ...................................................................................................................... 11
Host Functions (Raspberry Pi 3A+) ................................................................................. 11
Klipper Host Firmware ............................................................................................... 11
Mainsail Web GUI ...................................................................................................... 11
Wi-Fi Networking ....................................................................................................... 11
GPIO-Based Fan Control ............................................................................................ 11
Remote Camera Monitoring ....................................................................................... 11
ADXL345 (Input Shaper Accelerometer) ...................................................................... 12
MCU Functions (Ender 3 v2 Mainboard) .......................................................................... 12
Reflections ................................................................................................................... 12
Stability .................................................................................................................... 12
More Information ...................................................................................................... 12
6.Electrical Enclosure Design & Construction .................................................................... 12
Initial Design ................................................................................................................ 13
Electrical Enclosure Design ........................................................................................... 14
Design Goals ............................................................................................................ 14
Implementation Challenges ....................................................................................... 14
Considerations: ........................................................................................................ 14
Integration ................................................................................................................ 15
Reflections ................................................................................................................... 15
Improvements .......................................................................................................... 15
7. Electrical Systems ........................................................................................................ 16
Overview ...................................................................................................................... 17
Design Goals ................................................................................................................ 17
Reflections ................................................................................................................... 17
Improvements .......................................................................................................... 17
8. Software Setup & Automation ........................................................................................ 18
iii Table of Contents
Klipper & MCU Overview................................................................................................ 18
Why Klipper over Marlin ............................................................................................. 18
MCU & Klipper Firmware ............................................................................................ 18
Mainsail Frontend ......................................................................................................... 19
Fan Control and Macros ................................................................................................ 19
Design Goals ............................................................................................................ 19
Custom Fan Logic ..................................................................................................... 20
Start and End G-code .................................................................................................... 20
Configuration File ......................................................................................................... 20
Reflections ................................................................................................................... 21
Klipper Fan Logic Constraints..................................................................................... 21
9. Print Quality & Testing ................................................................................................... 21
Overview ...................................................................................................................... 21
Build Plate Probing ....................................................................................................... 21
Overview .................................................................................................................. 21
Skew Correction........................................................................................................ 22
Topology Compensation ............................................................................................ 22
Setup Considerations ................................................................................................ 23
Topographic Probing Test Methodology....................................................................... 23
Test Results .............................................................................................................. 23
Pressure Advance ......................................................................................................... 24
Overview .................................................................................................................. 24
Setup Consideration ................................................................................................. 24
Test Methodology ...................................................................................................... 24
Test Results .............................................................................................................. 25
Input Shaper ................................................................................................................ 25
Overview .................................................................................................................. 25
Test Methodology ...................................................................................................... 25
Manual Tuning .......................................................................................................... 26
Automatic Tuning ...................................................................................................... 26
Test Results .............................................................................................................. 26
10. Final Reflections & Future of the Build .......................................................................... 27
iv Table of Contents
Appendices ..................................................................................................................... 28
Appendix A: Print Head Shroud ........................................................................................ A
Appendix B: Print Head Explode View ............................................................................... 2
Appendix C: Duct CFD Analysis ....................................................................................... 3
Appendix D: Electrical Enclosure Explode-View ................................................................ 4
Appendix E: Electrical Schematic (KiCad) ......................................................................... 5
Appendix F: Electrical Enclosure Components ................................................................. 6
Appendix G: Fan Control Logic Flowchart ......................................................................... 7
Appendix H: X-Axis Input Shaper Graph ............................................................................ 8
Appendix I: Y-Axis Input Shaper Graph .............................................................................. 9
Appendix J: Rear Vent Technical Drawing. ....................................................................... 10
Appendix Y: Start & End G-Code..................................................................................... 11
Appendix Z: Printer Configuration .................................................................................. 12
Source and Modifications .......................................................................................... 12
Summary .................................................................................................................. 12
1 Portfolio Summary
Portfolio Summary
Figure 1: Shows enhanced printer with component list.
This project documents what started as a stock Ender 3 V2 3D printer and how I used it to
expand and refine my skills through hands on engineering.
Core Skill Highlight
• Firmware configuration and G-code automation (Klipper)
• Parametric CAD modelling and constraint-based design (Onshape)
• Sensor-driven motion tuning (Input Shaping, Pressure Advance)
• Automated build plate probing and real-time mesh compensation
• CFD-analysed airflow design and GPIO-based fan control
• Modular electrical rewiring with signal-integrity planning
• Multi-system integration: mechanical, electrical, and software
• Remote monitoring and diagnostics (Raspberry Pi, Mainsail)
• Iterative problem-solving under real-world constraints
2 1. Introduction
1. Introduction
I have always been drawn to complex systems and problem solving. Whether converting RC
cars into boats as a child, training as an electrician, or building and maintaining PCs, I’ve
looked for ways to take things apart, understand how they work, and improve them.
3D printing felt like a natural extension of that. The idea of being able to create obscure parts
that would’ve otherwise taken hours of searching or bodging together was especially
appealing. It felt like the most direct way to bring my imagination to fruition.
Before the printer even arrived, I had already started teaching myself CAD because I wanted to
do more than just print models – I wanted to design them.
What began as a basic machine turned into a launchpad for skills in CAD, CFD, firmware, low-
voltage electrical systems, motion mechanics, and real-world problem solving.
I’ve tackled everything from airflow tuning with CFD analysis and GPIO scripting to structural
vibration and thermal management.
Along the way, I have come to understand not just how the machine works, but how to make it
better.
This document is not a step-by-step guide, nor does it cover every minutia of the design
decisions that were made. It is a thoughtful reflection on the work, the challenges, and the
engineering thinking behind some key decisions - what worked, what didn’t, and what I would
improve.
3 2. Print Head & Cooling System
2. Print Head & Cooling System
Figure 2: Shows enhanced print head.
Figure 3: Shows explode view of enhanced print head assembly.
(See Appendix A for a multi-view of the print head shroud)
(See Appendix B for large format explode-view of the print head shroud)
4 2. Print Head & Cooling System
Overview
Having already designed several models of my own, including the electrical enclosure, my
intention was to try out a pre-existing shroud as a temporary enhancement before making my
own. However, that quickly turned into a major overhaul, mainly because I liked the
fundamentals of the design and saw clear opportunities to improve it.
The original design was appealing due to its low part count and reuse of stock components,
including the (40×10mm) axial (heatbreak) heatsink fan and a (20×10mm) centrifugal part
cooling fan. It offered a way to evaluate a new setup without any upfront investment, while still
improving on the baseline configuration.
What follows is a breakdown of the original design’s limitations, how my plan evolved, and the
key changes I made to expand its scope – particularly in airflow dynamics, whilst maintaining
structural stiffness.
Ducting
Figure 4: Shows CFD comparison of ducting. Enhanced(Left). Original (Right).
(See Appendix C for larger presentation).
Overview
The goal was quieter but more powerful part cooling with full Klipper/G-code control of the fan.
To achieve this, I increased the inlet size and duct volume, improving the flow paths to reduce
backpressure and support a larger PWM-controlled (50×20mm) centrifugal fan.
5 2. Print Head & Cooling System
Pre-CFD
The original design was allegedly CFD-evaluated, so I assumed that both nozzles were already
airflow-balanced. To avoid disrupting that balance, I mirrored any structural changes made to
one side of the duct on the other.
While identical geometry changes wouldn’t guarantee equal nozzle airflow, especially with
asymmetries in duct length and corners. It provided a reasonable basis for assuming
predictable flow behaviour.
Post-CFD
I’ve since validated this assumption through CFD simulation in SimScale. The results
confirmed that the mirrored design changes maintain a balanced airflow across both
channels, with only slight divergence due to expected path resistance differences.
Part Cooling Fan Efficiency
One issue I noticed early on was that the original fan orientation would have blocked one of the
upgraded fan's intakes due to how compactly it was mounted. While the obvious route would
have been to keep everything tightly packed, minimising overhangs and reducing lever arms
that could introduce ringing – I leaned toward improving airflow performance instead.
To ensure both intakes remained fully open, I rotated the fan 90 degrees and continued
splitting the airflow into two separate channels – accounting for the higher resistance on one
side due to its added length and more corners.
6 2. Print Head & Cooling System
ADXL345 Mounting
Figure 5: Show ADXL345 mount.
Overview
A mount was added to the print head to facilitate the use of Input Shaper on the X-axis.
The mounting point was chosen for its assumed stiffness, giving the accelerometer the
cleanest signal, and because it allows it to remain in its default orientation, eliminating the
need for additional configuration in Klipper.
More Information
For details on Input Shaper, see Section 9: Print Quality & Testing/Input Shaper.
7 2. Print Head & Cooling System
Larger Fan Mount
Figure 6: Shows part cooling fan mount.
Overview
The part cooling fan mount needed to be secure enough to reduce ringing, but still serviceable.
By making the mount a separate part, I could replace the fan without needing to reprint the
entire shroud.
Reflections
Starting from Scratch
My changes leaned toward stiffness over minimal weight, in part because Input Shaping deals
better with mass than with parts that flex or twist under inertia.
Looking back, I could have started from scratch and designed a new shroud entirely in
Onshape.
Because the original model was not parametric, making changes was laborious. I had to strip
away fillets just to move surfaces and spent hours visually inspecting and making more
changes.
The original model also didn’t fit my variation of mounting plate, so that had to be fixed.
That said, I do not regret sticking with it. Part of the value was in adapting someone else’s
model under constraints and I learnt a great deal from working within those boundaries.
8 3. Dual Z-Axis Synchronisation
Future Improvements
Strain Relief
I would also build in a proper strain relief system for cables, rather than relying on zip ties.
Water Cooling
For future enclosed setups, I have acquired a water-cooled heat break. It will allow me to eject
heat-break thermal energy more effectively, regardless of chamber temperatures and help
reduce fan noise.
3. Dual Z-Axis Synchronisation
Figure 7: Shows Dual Z-Axis system with synchronisation belt and tensioner.
Overview
This enhancement was necessary to improve the dimensional accuracy of printed parts.
On the stock Ender 3 V2, the Z-axis lead screw only supports the gantry on one side, which
allows the unsupported side to sag slightly over time. This results in layer misalignment and
poor first-layer consistency.
Before this enhancement, I avoided most Z-banding issues by manually adjusting the
unsupported side of the gantry slightly higher and allowed it to settle. It worked well enough to
avoid print defects, but it was a hassle to maintain and prone to drift over time, especially after
gantry bumps or manual movement.
9 3. Dual Z-Axis Synchronisation
Solution
I initially purchased a kit that included, a stepper motor, a lead screw, and a simple parallel
splitter cable – that connects the two stepper motors in series.
Electrically, the motors are not independently controlled. This is a completely passive system.
This kit also provides no physical link between them, so synchronisation over time cannot be
guaranteed.
To fix this, I added a mechanical synchronisation system that included matching pulleys on
each Z-axis leadscrew, that’s connected with a continuous timing belt.
This ensures that both sides of the gantry stay level, even after power loss or manual
movement.
Reflections
Tensioner Issue
This enhancement was mechanically simple in theory, but I didn’t fully account for real-world
tolerances, such as the belt arriving too long. I had calculated a “perfect” belt length based on
pulley pitch and spacing, aiming for a zero-slack system.
While I’m familiar with typical manufacturing variance, I now realise that assuming a perfect fit
left no margin for adjustment. In hindsight, I should have planned for an adjustable range from
the beginning.
To correct the mismatch, I retrofitted a tensioner to bring the belt into spec. It works well, but
the preload adjustment is fixed and a little awkward to modify.
That said, adding a tensioner wasn’t just a workaround – it’s a sound design choice. It allows
for correct belt tension without much work, making the system more adaptable and easier to
fine-tune.
The mechanical synchronisation system itself has proven reliable. It eliminated the need for
manual gantry correction and made the Z-axis far more robust and repeatable over time, even
after power loss or mechanical disturbance.
Future Improvements
Tensioner
In the future, I would incorporate adjustable tensioning from the outset and treat “perfect fit”
as a tolerance range, not a hard target.
10 4. Filament Drying System
4. Filament Drying System
Overview
Before I acquired a dedicated dryer, I assumed vacuum-sealed filament came pre-dried. I was
wrong.
Early PETG prints were riddled with bubbling and stringing, obvious signs of excessive moisture
in the filament. It quickly became obvious that I needed better moisture control, especially
when printing hygroscopic plastics.
Initial Solution
My first attempt was using a conventional oven. I added a thermocouple to monitor the
temperature and used a PC fan for basic air recirculation.
This setup worked surprisingly well but had several major drawbacks. It was energy-inefficient,
monopolised the kitchen oven, and made filament drying a chore rather than a background
process.
Solution
To resolve these issues, I purchased the Creality Filament Dry Box 2.0. It’s nothing fancy and
has a few issues that I’ll outline below, but it provides a dedicated and space-efficient
alternative to the oven.
Reflections
Thermal Performance
Beyond moisture control, preheating filament closer to its glass transition temperature
reduces the thermal load on the hotend. This can slightly increase maximum flow rate by
improving extrusion efficacy.
However, overheating can soften filament prematurely or degrade its properties, so control is
critical.
Ventilation Limitations
The dryer lacks active ventilation so over time, the chamber saturates with humid air, stalling
further evaporation. My workaround has been to manually open the lid every few hours during
long drying sessions to vent moisture buildup.
Temperature Stratification
Thermocouple tests confirmed significant temperature gradients, in the drying chamber. The
air near the heater is noticeably hotter than the air near the top. This leads to uneven drying
across the spool.
11 5. Host & MCU
To address this, I manually rotate the spool to even out heat exposure when not printing.
(Printing naturally rotates it.)
Current Material Limitations
Currently, I have only printed with PETG and PLA.
Higher-performance materials like Nylon and ABS are not practical with this machine due to
the 250°C nozzle temperature limit and the lack of a heated chamber to prevent warping.
Future Improvements
Ventilation
Automatic ventilation based on internal humidity.
Spool rotation
A low-RPM spool rotation system to ensure even drying across the roll.
5. Host & MCU
Overview
A host device was added because, even with custom firmware the Ender 3 V2’s stock
motherboard (MCU) is heavily constrained by its processing power and expandability. To
offload computation and enable advanced features, I added a Raspberry Pi 3A+ as a dedicated
host.
Host Functions (Raspberry Pi 3A+)
Klipper Host Firmware
The core of Klipper and interfaces with the MCU via USB.
Mainsail Web GUI
Provides a lightweight, browser-based interface for monitoring and control.
Wi-Fi Networking
Manages all wireless communication, enabling remote access and control.
GPIO-Based Fan Control
Uses Raspberry Pi GPIO pins to control fans and read their state.
(For details on Fan Control, see Section 8: Software Setup & Automation.)
Remote Camera Monitoring
A Pi camera module provides real-time print monitoring via the web interface.
12 6. Electrical Enclosure Design & Construction
ADXL345 (Input Shaper Accelerometer)
Connects via the GPIO pins to offer resonant frequency compensation.
(For details on Input Shaper, see Section 9: Print Quality & Testing.)
MCU Functions (Ender 3 v2 Mainboard)
The stock Ender 3 V2 mainboard remains in use. It is flashed with Klipper firmware and now
manages the topographic build plate probe and real-time stepper motion execution, with all
high-level processing offloaded to the host.
Reflections
Stability
This setup has been stable, with no crashes or hangups across days and weeks of print time.
The camera system has also worked without issue.
More Information
For details on software configuration, including Klipper setup, fan control logic, and macros,
see Section 8: Software Setup & Automation.
For detail on advanced testing features, including Topographic Build Plate Probing, Pressure
Advance and Input Shaper, see Section 9: Print Quality & Testing.
6. Electrical Enclosure Design & Construction
Figure 8: Shows transparent CAD model of the electrical enclosure.
13 6. Electrical Enclosure Design & Construction
Figure 9: Shows explode view of the electrical enclosure.
(See Appendix D for a larger presentation).
Initial Design
The stock printer cooling intakes are located on its underside, less than 2 cm from the surface
it sits on.
Due to those airflow restrictions, I had designed ducts to route airflow upward through the
mounting surface. But soon after I finished modelling the parts, I realised I was trying to
engineer around what I thought was a fundamentally bad design. Even if airflow had a direct
path, the fans were still too small and loud.
14 6. Electrical Enclosure Design & Construction
Figure 10: Shows through table grommet design with stock printer.
The initial grommet design intended to improve airflow. Although this was not implemented,
the exercise informed my final electrical enclosure redesign.
Electrical Enclosure Design
After scrapping the original cooling system and its intended fix, I set out to redesign the
printer’s electrical enclosure from the ground up.
Design Goals
• Compact footprint for an efficient use of space
• Larger fan support for higher airflow and less noise
• Housing for Raspberry Pi, buck converter, and printer’s mainboard
• Secure internal and external mounting for all components and cables
• Enhanced serviceability
Implementation Challenges
At face value, none of these goals are particularly complicated but integrating them all into one
coherent design turned out to be one of the hardest parts of this entire build.
Considerations:
• Physical tolerances and alignment
• Cable runs and connector accesses
• Internal airflow routing
• Electromagnetic interference
15 6. Electrical Enclosure Design & Construction
Integration
It is what people sometimes call "integration hell", fixing one thing would often break
something else.
Thankfully, it wasn’t too much of an issue. Onshape’s parametric modelling really helped me
during this phase.
It let me keep dependencies connected and respond to changes without breaking the model.
Reflections
This electrical enclosure was the culmination of everything I’ve learnt in CAD. It’s the most
iterated design in this whole project, and though it was frustrating at times, learning how to
build something where multiple systems meet; and finding a way to make it all work is just part
of why I love engineering.
Even though the result is a marked improvement over the stock configuration, there are still
things I would like to improve.
Improvements
External Connectors
Routing all power and communication through external connectors instead of direct
passthroughs would make the system much easier to maintain or reconfigure.
Electromagnetic interference
Even though I moved the buck converter away from the Raspberry Pi's GPIO pins, I'd like to
design EMF mitigation into the electrical enclosure.
Reinforcement
Increase the electrical enclosure infill and reinforce mounting points to the aluminium
extrusion.
PSU & Electrical Enclosure Ducting
Enclose the PSU in a dedicated housing and add ducting for both it and the electrical
enclosure. This would allow external airflow to be routed in and out without affecting internal
chamber temperatures, helping to prevent print warping if the printer is ever used in a heated
chamber.
This is important because internal ambient temperatures could reach 40–50 °C, in a heated
chamber, which would be too high for reliable operation without dedicated airflow.
Despite these areas for improvement, the electrical enclosure achieved all major design goals.
The airflow is more efficient thanks to the larger fan and the less restrictive internal layout.
16 7. Electrical Systems
7. Electrical Systems
Figure 11: Shows electrical schematic of Ender 3 v2 after enhancement.
(See Appendix E for a larger presentation).
Figure 12: Shows photograph of the components in the electrical enclosure.
17 7. Electrical Systems
Overview
The stock Ender 3 V2 wiring loom is mostly adequate, but its biggest weakness is
serviceability.
In the event of a failure, any work would have to be conducted with the printer on its side, as all
terminations are made inside the original printer’s cramped electrical enclosure on the
underside, with no alternative access.
While I would have liked to rewire the loom entirely with proper modular terminations, that was
outside the scope.
Design Goals
• Space efficient custom cable for Host-to-MCU communication.
• JST connecters for the part cooling fan to improve serviceability.
• Longer 24V main power cable with buck converter spur.
• Ethernet cable twisted pairs connect ADXL to Host to aid with signal integrity.
• Tune Vref on MCU to lower stepper motor temperature while maintaining torque.
Reflections
This part of the project was straightforward, but it could grow significantly more complex in the
future.
Improvements
Print Head Daughterboard
Supplying a single 24V line to the print head with onboard 12V, 5V, and 3.3V regulation for any
fans, sensors, or logic. This would reduce the amount of power lines, which run to the print
head.
Using CAN bus for all data communication, providing a reliable and condensed data path that
would also simplify the wiring loom.
External Termination
The most important thing I would change in hindsight is to add external terminals to all major
connections – especially between the print head and electrical enclosure.
Right now, every cable replacement requires cutting zip ties in the electrical enclosure and re-
routing cables, which makes even simple maintenance a hassle.
I had originally considered reusing an Ethernet jack for the ADXL345, but I was not comfortable
with the small risk of someone accidentally plugging in a live network cable.
In the future, I will add a small USB-powered MCU like a Raspberry Pi Zero to function as a
signal bridge.
18 8. Software Setup & Automation
That way, even if someone plugs in another device, nothing gets damaged.
8. Software Setup & Automation
Klipper & MCU Overview
Custom firmware like Klipper was vital to the success of this project.
It allows rapid iteration of configuration settings and hardware changes without the need to
manually recompile the firmware, and supports advanced features such as flow rate tuning
and harmonic compensation.
Why Klipper over Marlin
For many enthusiasts, the capabilities mentioned have long given Klipper an edge over Marlin,
and they’re part of the reason many newer consumer machines now ship with Klipper
derivatives by default.
While Marlin has begun to incorporate some similar features, it still doesn’t match Klipper’s
flexibility or ease of customisation.
MCU & Klipper Firmware
Klipper offloads all computational workload from the MCU to a host device – allowing the MCU
to focus solely on executing motion instructions for the stepper motors and deploying the
topographic build plate probe, reading the probe state, and retracting it.
This architecture eliminates the performance bottlenecks of a low-power control board.
As a result, I saw no practical reason to upgrade the Ender 3 V2’s stock mainboard. It performs
well within this distributed setup.
19 8. Software Setup & Automation
Mainsail Frontend
Figure 13: Shows Mainsail dashboard.
Mainsail was chosen for the frontend due to its active development, responsive UI, and strong
integration with Klipper’s configuration model.
It provides real-time control, temperature graphing, print monitoring, and command input
through a web browser on the local network.
This combination, along with a Raspberry Pi camera module, offers full remote access and
monitoring. This is critical for checking first-layer adhesion and print progress whilst away from
the printer.
Fan Control and Macros
Using Klipper’s config system, G-code macros, and the Pi’s GPIO support, I set out to
implement dynamic fan control and general automation behaviour.
Design Goals
• Custom activation thresholds
• RPM-based feedback via tachometers
• Custom start and end print macros
• Live monitoring via Mainsail
20 8. Software Setup & Automation
Custom Fan Logic
Electrical Enclosure Fan
Activates when stepper motors are enabled and or Raspberry Pi CPU reaches 50°C; turns off
after idle timeout if Raspberry Pi CPU is below 50°C
Heat Break Fan
Turns on when the hotend exceeds 50 °C; Turns off if below 50°C
Figure 14: Shows custom fan logic flowchart.
(See Appendix G for large format of fan logic flowchart)
Start and End G-code
Start and end sequences are managed through Klipper macros rather than slicer settings.
This avoids redundant maintenance across multiple slicers and centralises configuration.
• START_PRINT macro; Lays down two purge lines to prime the hotend and clear any
degraded plastic.
• END_PRINT macro; Performs a long retraction to reduce residual pressure and help
prevent clogging or oozing during cooldown.
(See Appendix Y for full start and end G-code.)
Configuration File
See Appendix Z for full printer configuration, including motion tuning parameters and GPIO fan
control setup.
21 9. Print Quality & Testing
Reflections
Klipper Fan Logic Constraints
Klipper’s fan control logic has some constraints – for example, it does not allow a single fan to
respond to multiple independent triggers.
Because of that, I chose to prioritise stepper driver cooling over Pi CPU temperature, as the
CPU stays under 50 °C passively and is less sensitive to heat buildup.
Though Klipper recommends hardware PWM in high-load environments to avoid timing jitter
from software-based control. In practice, software PWM has been completely dependable in
my setup, due to the Pi’s light computational load.
Aside from that, the macros and fan behaviours have worked well. I am especially happy with
how centralised and configurable this setup has been.
9. Print Quality & Testing
Overview
Due to their complex kinematics, varied part selection, and the performance of individual
components, 3D printers require some form of calibration.
As print speeds increase, this calibration becomes more involved and may require sensor-
based real-time feedback loops to maintain quality.
This is not an exhaustive list of every test used, but I will outline the most notable ones in my
printer’s setup.
The three that I will go over are Topographic Build Plate Probing, Pressure Advance and Input
Shaping, the last two of which are mandatory for high-speed printing. I will briefly explain how I
interpret their function, how the tests were conducted, and how I validated the results.
Build Plate Probing
Overview
The BLTouch replaces the Z-axis end-stop and adds two critical functions that allow the nozzle
to maintain the optimal distance from the build plate throughout printing.
22 9. Print Quality & Testing
Figure 15: Shows bed probing operation and some of the components involved.
Skew Correction
To correct for skew across the X and Y axes, the probe samples all four corners of the bed and
reports how far each point is from the home reference (Z = 0). These offsets are then used to
guide manual adjustment of the bed levelling screws, usually via corner-mounted dials.
Klipper makes this easier by factoring in the Z-offset, thread pitch, and dial diameter, then
outputting precise adjustment instructions, often in a clock-face format (e.g. “turn
anticlockwise 10 minutes”).
Topology Compensation
Once skew is minimised, the probe samples a grid of points across the usable build surface,
building a topographic mesh. Klipper then applies real-time Z-offset adjustments during
printing to account for surface irregularities.
This enables consistent nozzle height across warped or uneven areas, which improves first-
layer reliability, reduces the need for manual tuning, and enhances print dimensional
accuracy.
Klipper also supports multiple saved meshes. This is useful when compensating for thermal
expansion at different build plate temperatures.
23 9. Print Quality & Testing
Setup Considerations
To avoid interference from bed bolts or clips, the probing area is constrained within a user-
defined “keep-in” region. The resolution of the mesh is configurable in firmware, but probing
time increases with density. For this printer, a 5×5 grid provides a good balance between
accuracy and efficiency, as the bed is relatively flat at this scale.
Topographic Probing Test Methodology
To evaluate the mesh compensation, I used a single-layer flatness test print.
This consists of a large grid pattern that covers the full build surface. The goal is even
extrusion across all regions, without areas of over-squish or nozzle-to-bed contact.
I also manually adjusted the Z-offset after probing to fine-tune the first-layer height.
Test Results
Figure 16: Shows height map of build plate.
Using a 5×5 probing grid in Mainsail, the mesh revealed approximately 0.1 mm of total
deviation across the bed. The highest point was near the front-left corner (+0.092 mm), and the
lowest was slightly behind and to the right of centre (–0.007 mm).
After enabling mesh compensation in Klipper, first-layer test prints showed consistent and
even extrusion across the full build surface, confirming both the accuracy of the mesh and the
mechanical stability of the system.
24 9. Print Quality & Testing
Pressure Advance
Overview
Pressure Advance works by pre-emptively extruding or retracting filament based on upcoming
motion. Anticipating what the nozzle is about to do rather than reacting in real time.
The importance of this tool becomes more obvious as print speeds increase. As the print head
moves faster, the delay between commanded and actual extrusion becomes more noticeable
– especially during rapid directional changes and at seam junctions.
An example is when the extruder is printing a sharp right angle, as the nozzle arrives at the
corner, it momentarily pauses. Due to residual pressure in the system, filament keeps flowing,
resulting in over-extrusion on the corner. The same logic applies as the nozzle exits the corner.
Pressure Advance counteracts this by slightly retracting the filament before reaching the
corner, reducing flow just in time for the slowdown.
Without compensation, the sudden acceleration and delay in flow would initially cause under-
extrusion.
Pressure Advance solves this by commanding extra flow just before it is needed, effectively
smoothing the transition.
Setup Consideration
One key consideration is that Pressure Advance needs to be calibrated for every filament type,
brand, and print temperature, since it is highly dependent on material viscosity.
This is where ecosystems with proprietary filament shine, the manufacturer has often done
most of the tuning work and the printer can apply the settings for you.
Test Methodology
Tuning Pressure Advance is straightforward.
All that was needed was a custom-sliced test object (usually a simple tower), a few lines of G-
code, and a pair of vernier callipers.
The G-code instructs the printer to incrementally adjust the Pressure Advance value at each
layer.
Once the print is complete, the callipers are used to find the point on the tower where print
quality is the cleanest, without bulging or gaps.
25 9. Print Quality & Testing
Test Results
Figure 17: Shows pressure advance calibration tower with measurement example.
Input Shaper
Overview
To further improve motion accuracy, I used Klipper’s Input Shaping feature. Which is enabled
using an accelerometer (ADXL345).
The goal here is to reduce resonant vibrations caused by frame movement, belt tension, and
the inertia of moving parts.
Input Shaper works by identifying dominant resonance frequencies in the motion system and
applying an equal and opposite force, cancelling out the unwanted vibration.
Think of Newton’s third law. Every action has an equal and opposite reaction, so If both
opposite forces are applied to the same structure, the motion cancels out.
Of course, it is not a perfect system. Real-world physics introduces signal noise, frame flex,
and other chaos but the results are impressive, especially on less rigid machines.
Test Methodology
There are two ways to tune Input Shaper in Klipper
26 9. Print Quality & Testing
Manual Tuning
This requires manual visual pattern analysis to find the frequency nodes. It works well but less
accurate and thus results will be limited.
Automatic Tuning
This uses direct measurement from an accelerometer.
I used the latter because it is faster, far more accurate, and produces a clean dataset for
comparison.
The first step is to measure the signal-to-noise ratio (SNR) of the accelerometer because a
clean signal ensures more accurate data. Noise in = noise out.
Once verified, the sensor is mounted on each axis.
The test is then run for the X-axis (print head carriage) and Y-axis (build plate).
The Z-axis is excluded from this process since Cartesian printers only move that axis during
layer changes or Z-hops – it doesn’t contribute to system resonance during normal printing.
Test Results
Figure 18: Show Input Shaper calibration results (X & Y Axes).
(See Appendix H and I for larger graphs of Input Shaper test results).
Resonance analysis was performed using an ADXL345 sensor mounted on the toolhead and
Build-plate.
27 10. Final Reflections & Future of the Build
The X-axis (toolhead) showed a peak at 67.4 Hz and was tuned using the MZV shaper.
The Y-axis (build plate) showed a much lower peak at 35.2 Hz, also tuned with MZV.
To prevent over-smoothing, maximum acceleration was capped at 3700 mm/s², as
recommended.
10. Final Reflections & Future of the Build
This project was the result of long nights and days spent thinking about CAD.
It gave me a platform to learn new skills and stretch my problem-solving muscles. I often found
myself daydreaming about how to improve a part, constantly iterating on ideas, even when I
was away from the printer.
Honestly, finishing the build was bittersweet because I enjoyed the process more than I
expected. It made me realise that this kind of work is something I want to pursue more
seriously.
As for the printer itself, it has reached the limits of what makes sense to modify. With enough
tuning, I could squeeze out a bit more performance but like that very first duct I designed, I
would just be engineering around the platform’s limitations.
The motion system lacks stiffness, so motion accuracy suffers at higher speeds. And if I am
honest, print quality is not hugely better than stock. Though it was particularly good to begin
with and it is much easier to achieve consistent results when printing slowly.
That said, the machine is now faster, more capable, and far more serviceable than it ever was
out of the box.
Looking ahead, I hope to move to a CoreXY printer, ideally something open source so I can
keep full control and continue building on what I have learned here.
Finally, if I may indulge for a moment.
This project was about more than just 3D printing. Yes, it helped me in practical ways, and
gave me a platform to experiment, refine, and grow.
But more than that, it gave me something solid to hold onto during a time when I really needed
it – something that challenged, absorbed, and reminded me what I’m capable of when I follow
my curiosity.
It’s a technical document, but it’s also a human story. One that’s about learning through
doing, and finding focus and meaning by building something I genuinely care about.
It is the best £200 I’ve ever spent.
Thank you for reading.
Appendices
Supporting Diagrams, Photographs, Macros, and Configuration Files
A Appendices
Appendix A: Print Head Shroud
Figure 19: Shows print head shroud. (Enhanced Top. Original Bottom).
B Appendices
Appendix B: Print Head Explode View
Figure 20: Shows explode view of print head shroud.
C Appendices
Appendix C: Duct CFD Analysis
Figure 21: Shows SimScale CFD analysis of the enhanced and original print head shroud.
D Appendices
Appendix D: Electrical Enclosure Explode-View
Figure 22: Shows explode view of assembled the electrical enclosure.
E Appendices
Appendix E: Electrical Schematic (KiCad)
Figure 23: Shows enhanced printers electrical schematic.
F Appendices
Appendix F: Electrical Enclosure Components
Figure 24: Shows photograph of the components in the electrical enclosure.
G Appendices
Appendix G: Fan Control Logic Flowchart
Figure 25: Shows fan logic for the heat break and electrical enclosure.
H Appendices
Appendix H: X-Axis Input Shaper Graph
Figure 26: Shows Input Shaper test results for X and Y axis of the print head.
I Appendices
Appendix I: Y-Axis Input Shaper Graph
Figure 27: Shows Input Shaper test results for Y-axis of the build plate.
J Appendices
Appendix J: Rear Vent Technical Drawing.
Figure 28: Shows technical drawing of the rear vent.
K Appendices
Appendix Y: Start & End G-Code
[gcode_macro START_PRINT]
gcode:
# Start bed heating
#M140 S75
# Use absolute coordinates
G90
# Reset the G-Code Z offset
(adjust Z offset if needed)
SET_GCODE_OFFSET Z=0.0
# Home the printer
G28
# Wait for bed to reach
temperature
#M190 S75
BED_MESH_PROFILE LOAD=default
# Set and wait for nozzle to
reach temperature
#M104 S250
#M109 S250
# Move to wait position
G1 X0 Y0 Z30 F4000.0
# Move Z-axis up
G1 Z2.0 F3000
# Move to start position
G1 X10.1 Y20 Z0.28 F5000.0
# Reset extruder
G92 E0
# Draw the first line
G1 X10.1 Y200.0 Z0.28 F1500.0 E15
# Move to the side
G1 X10.4 Y200.0 Z0.28 F5000.0
# Draw the second line
G1 X10.4 Y20 Z0.28 F1500.0 E30
# Reset extruder
G92 E0
[gcode_macro END_PRINT]
gcode:
# Finish moves
M400
# Set positioning to relative
G91
# Retract the filament by 18 mm
G1 E-18 F800
# Retract the filament more and
raise Z-axis
G1 E-2 Z0.2 F2400
# Set positioning to absolute
G90
# Move the bed forward
G1 X0 Y220 F1000
# Turn off the hotend
M104 S0
# Turn off the bed
M140 S0
# Turn off the fan
M106 S0
# Disable steppers
M84
M84
L Appendices
Appendix Z: Printer Configuration
Source and Modifications
This configuration file began as Klipper’s official printer-creality-ender3-v2-2020.cfg template and
was adapted for this custom Ender 3 V2 rebuild. Many parameters were modified for hardware
compatibility and performance tuning, while others were added to support advanced features such
as input shaping and mesh levelling.
Section / Parameter
Klipper Default Value
Custom Value or Addition
position_max (X)
235
240
position_max (Y)
235
230
position_min (Z)
–5
–2.176
nozzle_diameter
0.400
0.600
pressure_advance
Not present
0.6
[bltouch]
Not present
Added (sensor setup and offsets)
[safe_z_home]
Not present
Added (BLTouch homing setup)
[bed_mesh]
Not present
Added (5×5 grid)
[screws_tilt_adjust]
Not present
Manual bed levelling configuration
[fan], [controller_fan]
Not present
GPIO-controlled fans + tachometer
[heater_fan Heatbreak_Fan]
Not present
Fan tied to extruder temperature
[adxl345], [resonance_tester]
Not present
Added for vibration testing
[input_shaper]
Not present
Added (based on calibration results)
Summary
This configuration reflects a tailored firmware setup that retains the structure of Klipper’s defaults
while layering on targeted modifications to accommodate new hardware, improve print quality, and
enable advanced features. The file has evolved iteratively through testing and calibration during the
rebuild.
M Appendices
# This file contains pin mappings for the stock 2020 Creality Ender 3
# V2. To use this config, during "make menuconfig" select the
# STM32F103 with a "28KiB bootloader" and serial (on USART1 PA10/PA9)
# communication.
# If you prefer a direct serial connection, in "make menuconfig"
# select "Enable extra low-level configuration options" and select
# serial (on USART3 PB11/PB10), which is broken out on the 10 pin IDC
# cable used for the LCD module as follows:
# 3: Tx, 4: Rx, 9: GND, 10: VCC
# Flash this firmware by copying "out/klipper.bin" to a SD card and
# turning on the printer with the card inserted. The firmware
# filename must end in ".bin" and must not match the last filename
# that was flashed.
# See docs/Config_Reference.md for a description of parameters.
[include mainsail.cfg]
[include macro.cfg]
[stepper_x]
step_pin: PC2
dir_pin: PB9
enable_pin: !PC3
microsteps: 16
rotation_distance: 40
endstop_pin: ^PA5
position_endstop: 0
position_max: 240
homing_speed: 50
[stepper_y]
step_pin: PB8
dir_pin: PB7
enable_pin: !PC3
microsteps: 16
rotation_distance: 40
endstop_pin: ^PA6
position_endstop: 0
position_max: 230
homing_speed: 50
[stepper_z]
step_pin: PB6
dir_pin: !PB5
enable_pin: !PC3
microsteps: 16
rotation_distance: 8
endstop_pin:
probe:z_virtual_endstop
#position_endstop: -0.1
position_max: 250
position_min: -2.176
[bltouch]
sensor_pin: ^PB1
control_pin: PB0
x_offset: -44
y_offset: -6
[safe_z_home]
home_xy_position: 117.5,117.5
z_hop: 10
z_hop_speed: 10
[bed_mesh]
speed: 120
horizontal_move_z: 5
mesh_min: 15,15
mesh_max: 188,191
probe_count: 5,5
algorithm: bicubic
fade_start: 1
fade_end: 10
fade_target: 0
[screws_tilt_adjust]
screw1: 74, 42
screw1_name: Front left
screw2: 235, 42 # Adjusted to
fit within the bed size
screw2_name: Front right
screw3: 235, 213 # Adjusted to
fit within the bed size
screw3_name: Back right
screw4: 74, 213
screw4_name: Back left
screw_thread: CW-M4
N Appendices
horizontal_move_z: 10
speed: 200
[extruder]
max_extrude_only_distance: 100.0
step_pin: PB4
dir_pin: PB3
enable_pin: !PC3
microsteps: 16
rotation_distance: 34.406
nozzle_diameter: 0.600
filament_diameter: 1.750
heater_pin: PA1
sensor_type: EPCOS 100K
B57560G104F
sensor_pin: PC5
#control: pid
# tuned for stock hardware with
200 degree Celsius target
#pid_Kp: 21.527
#pid_Ki: 1.063
#pid_Kd: 108.982
min_temp: 0
max_temp: 260
pressure_advance: 0.6
[heater_bed]
heater_pin: PA2
sensor_type: EPCOS 100K
B57560G104F
sensor_pin: PC4
control: pid
# tuned for stock hardware with
50 degree Celsius target
pid_Kp: 54.027
pid_Ki: 0.770
pid_Kd: 948.182
min_temp: 0
max_temp: 130
[temperature_sensor MCU]
sensor_type: temperature_mcu
[temperature_sensor RPI]
sensor_type: temperature_host
[fan]
pin: rpi:gpio12
max_power: 1.0
shutdown_speed: 0
cycle_time: 0.010
hardware_pwm: False
kick_start_time: 0.100
off_below: 0.0
tachometer_pin: ^rpi:gpio16
tachometer_ppr: 2
tachometer_poll_interval: 0.0015
[controller_fan Controller_Fan]
pin: rpi:gpio13
max_power: 1.0
shutdown_speed: 0.0
cycle_time: 0.010
hardware_pwm: False
kick_start_time: 0.100
off_below: 0.0
tachometer_pin: ^rpi:gpio26
tachometer_ppr: 2
tachometer_poll_interval: 0.0015
#enable_pin:
# See the "fan" section for a
description of the above
parameters.
fan_speed: 1.0
# The fan speed (expressed as a
value from 0.0 to 1.0) that the
fan
# will be set to when a heater
or stepper driver is active.
# The default is 1.0
idle_timeout: 30
# The amount of time (in
seconds) after a stepper driver
or heater
# was active and the fan should
be kept running. The default
# is 30 seconds.
idle_speed: 0.5
# The fan speed (expressed as a
value from 0.0 to 1.0) that the
fan
# will be set to when a heater
or stepper driver was active and
# before the idle_timeout is
reached. The default is
fan_speed.
heater: heater_bed, extruder
#stepper: stepper_x, stepper_y,
stepper_z, extruder
# Name of the config section
defining the heater/stepper that
this fan
# is associated with. If a
comma separated list of
heater/stepper names
# is provided here, then the
fan will be enabled when any of
the given
O Appendices
# heaters/steppers are enabled.
The default heater is "extruder",
the
# default stepper is all of
them.
[heater_fan Heatbreak_Fan]
pin: PA0
#max_power:
#shutdown_speed:
#cycle_time:
#hardware_pwm:
kick_start_time: 0.5
#off_below:
#tachometer_pin:
#tachometer_ppr:
#tachometer_poll_interval:
#enable_pin:
# See the "fan" section for a
description of the above
parameters.
heater: extruder
# Name of the config section
defining the heater that this fan
is
# associated with. If a comma
separated list of heater names is
# provided here, then the fan
will be enabled when any of the
given
# heaters are enabled. The
default is "extruder".
heater_temp: 50.0
# A temperature (in Celsius)
that the heater must drop below
before
# the fan is disabled. The
default is 50 Celsius.
#fan_speed: 1.0
# The fan speed (expressed as a
value from 0.0 to 1.0) that the
fan
# will be set to when its
associated heater is enabled. The
default
# is 1.0
[mcu]
serial: /dev/serial/by-id/usb-
1a86_USB_Serial-if00-port0
restart_method: command
[mcu rpi]
serial: /tmp/klipper_host_mcu
[adxl345]
cs_pin: rpi:None
[resonance_tester]
accel_chip: adxl345
probe_points: 117.5, 117.5, 20
[input_shaper]
#shaper_freq_x: 70.0
#shaper_type_x: mzv
#shaper_freq_y: 38.2
#shaper_type_y: mzv
[printer]
kinematics: cartesian
max_velocity: 300
max_accel: 3500
max_z_velocity: 25
max_z_accel: 500
#*# <----------------------
SAVE_CONFIG ---------------------
->
#*# DO NOT EDIT THIS BLOCK OR
BELOW. The contents are auto-
generated.
#*#
#*# [bltouch]
#*# z_offset = 2.350
#*#
#*# [bed_mesh default]
#*# version = 1
#*# points =
#*# 0.067500, 0.072500,
0.072500, 0.072500, 0.092500
#*# 0.057500, 0.045000,
0.037500, 0.037500, 0.047500
#*# 0.050000, 0.035000,
0.015000, 0.010000, 0.012500
#*# 0.067500, 0.042500,
0.025000, 0.005000, -0.007500
#*# 0.065000, 0.047500,
0.020000, 0.020000, 0.032500
#*# x_count = 5
#*# y_count = 5
#*# mesh_x_pps = 2
#*# mesh_y_pps = 2
#*# algo = bicubic
#*# tension = 0.2
#*# min_x = 15.0
#*# max_x = 188.0
#*# min_y = 15.0
#*# max_y = 191.0
P Appendices
#*#
#*# [extruder]
#*# control = pid
#*# pid_kp = 30.322
#*# pid_ki = 1.743
#*# pid_kd = 131.901
#*#
#*# [input_shaper]
#*# shaper_type_x = zv
#*# shaper_freq_x = 67.4
#*# shaper_type_y = mzv
#*# shaper_freq_y = 36.8
